Spatially Explicit Models of City Population Growth in India

Donghwan Kim, Yale University
Mark R. Montgomery, Population Council

This paper presents a state-of-the-art econometric model of city population growth in India using longitudinal data at the city level for a great number of cities and disaggregated time-series of important demographic determinants of this growth. In methodological terms, this paper combines panel data spatial econometrics with multi-level modelling. We depart from conventional models of spatial correlation in allowing spatial effects to emerge not only from geographic distance, but also from the political and economic context in which the city is situated, with important influences at the state and regional level in India as well as spill-over effects from neighboring cities. Our initial analysis shows that urban fertility rates display very strong positive effects on city population growth rates and confirms that city growth is spatially correlated. These results will be extended with more additional data and models in which spatial correlation stems from an explicitly multi-level structure.